public inbox for gcc-bugs@sourceware.org
help / color / mirror / Atom feed
* [Bug target/43766] New: x86 prefetch doesn't use complex memory addressing
@ 2010-04-16 7:52 astrange at ithinksw dot com
2010-04-16 11:37 ` [Bug target/43766] " ubizjak at gmail dot com
` (5 more replies)
0 siblings, 6 replies; 7+ messages in thread
From: astrange at ithinksw dot com @ 2010-04-16 7:52 UTC (permalink / raw)
To: gcc-bugs
Source:
void p(int *a, int i)
{
__builtin_prefetch(&a[i]);
}
> gcc -O3 -fomit-frame-pointer -S prefetch.c
_p:
movslq %esi, %rsi
leaq (%rdi,%rsi,4), %rax
prefetcht0 (%rax)
ret
leaq and prefetch should be merged.
--
Summary: x86 prefetch doesn't use complex memory addressing
Product: gcc
Version: 4.6.0
Status: UNCONFIRMED
Severity: enhancement
Priority: P3
Component: target
AssignedTo: unassigned at gcc dot gnu dot org
ReportedBy: astrange at ithinksw dot com
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=43766
^ permalink raw reply [flat|nested] 7+ messages in thread
* [Bug target/43766] x86 prefetch doesn't use complex memory addressing
2010-04-16 7:52 [Bug target/43766] New: x86 prefetch doesn't use complex memory addressing astrange at ithinksw dot com
@ 2010-04-16 11:37 ` ubizjak at gmail dot com
2010-04-16 18:07 ` ubizjak at gmail dot com
` (4 subsequent siblings)
5 siblings, 0 replies; 7+ messages in thread
From: ubizjak at gmail dot com @ 2010-04-16 11:37 UTC (permalink / raw)
To: gcc-bugs
------- Comment #1 from ubizjak at gmail dot com 2010-04-16 11:37 -------
Currently we fail to recognize address_operand in the form of:
(prefetch (plus:DI (ashift:DI (reg:DI 60 [ i ])
(const_int 2 [0x2]))
(reg/v/f:DI 58 [ a ]))
(const_int 0 [0x0])
(const_int 3 [0x3]))
This form should be added to the parser in ix86_decompose_address.
I'll look into it.
--
ubizjak at gmail dot com changed:
What |Removed |Added
----------------------------------------------------------------------------
AssignedTo|unassigned at gcc dot gnu |ubizjak at gmail dot com
|dot org |
Status|UNCONFIRMED |ASSIGNED
Ever Confirmed|0 |1
Last reconfirmed|0000-00-00 00:00:00 |2010-04-16 11:37:34
date| |
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=43766
^ permalink raw reply [flat|nested] 7+ messages in thread
* [Bug target/43766] x86 prefetch doesn't use complex memory addressing
2010-04-16 7:52 [Bug target/43766] New: x86 prefetch doesn't use complex memory addressing astrange at ithinksw dot com
2010-04-16 11:37 ` [Bug target/43766] " ubizjak at gmail dot com
@ 2010-04-16 18:07 ` ubizjak at gmail dot com
2010-04-16 21:19 ` astrange at ithinksw dot com
` (3 subsequent siblings)
5 siblings, 0 replies; 7+ messages in thread
From: ubizjak at gmail dot com @ 2010-04-16 18:07 UTC (permalink / raw)
To: gcc-bugs
------- Comment #2 from ubizjak at gmail dot com 2010-04-16 18:06 -------
Created an attachment (id=20402)
--> (http://gcc.gnu.org/bugzilla/attachment.cgi?id=20402&action=view)
Proposed patch.
Proposed patch generates:
p:
movslq %esi, %rsi
prefetcht0 (%rdi,%rsi,4)
ret
--
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=43766
^ permalink raw reply [flat|nested] 7+ messages in thread
* [Bug target/43766] x86 prefetch doesn't use complex memory addressing
2010-04-16 7:52 [Bug target/43766] New: x86 prefetch doesn't use complex memory addressing astrange at ithinksw dot com
2010-04-16 11:37 ` [Bug target/43766] " ubizjak at gmail dot com
2010-04-16 18:07 ` ubizjak at gmail dot com
@ 2010-04-16 21:19 ` astrange at ithinksw dot com
2010-04-17 15:11 ` ubizjak at gmail dot com
` (2 subsequent siblings)
5 siblings, 0 replies; 7+ messages in thread
From: astrange at ithinksw dot com @ 2010-04-16 21:19 UTC (permalink / raw)
To: gcc-bugs
------- Comment #3 from astrange at ithinksw dot com 2010-04-16 21:19 -------
Works with x86-64.
Checking -m32, the same thing happens with or without the patch:
_p:
subl $12, %esp
movl 20(%esp), %eax
sall $2, %eax
addl 16(%esp), %eax
addl $12, %esp
prefetcht0 (%eax)
ret
--
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=43766
^ permalink raw reply [flat|nested] 7+ messages in thread
* [Bug target/43766] x86 prefetch doesn't use complex memory addressing
2010-04-16 7:52 [Bug target/43766] New: x86 prefetch doesn't use complex memory addressing astrange at ithinksw dot com
` (2 preceding siblings ...)
2010-04-16 21:19 ` astrange at ithinksw dot com
@ 2010-04-17 15:11 ` ubizjak at gmail dot com
2010-04-19 12:37 ` uros at gcc dot gnu dot org
2010-04-19 12:38 ` ubizjak at gmail dot com
5 siblings, 0 replies; 7+ messages in thread
From: ubizjak at gmail dot com @ 2010-04-17 15:11 UTC (permalink / raw)
To: gcc-bugs
------- Comment #4 from ubizjak at gmail dot com 2010-04-17 15:11 -------
(In reply to comment #3)
> Works with x86-64.
>
> Checking -m32, the same thing happens with or without the patch:
This happens because combine pulls memory references into the combined address
operand. Since ix86_decompose_address expects registers everywhere, it fails.
The same happens with:
int * p(int *a, int i)
{
return &a[i];
}
So, this particular case is unfixable on -m32. Try with -mregparm=3.
--
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=43766
^ permalink raw reply [flat|nested] 7+ messages in thread
* [Bug target/43766] x86 prefetch doesn't use complex memory addressing
2010-04-16 7:52 [Bug target/43766] New: x86 prefetch doesn't use complex memory addressing astrange at ithinksw dot com
` (3 preceding siblings ...)
2010-04-17 15:11 ` ubizjak at gmail dot com
@ 2010-04-19 12:37 ` uros at gcc dot gnu dot org
2010-04-19 12:38 ` ubizjak at gmail dot com
5 siblings, 0 replies; 7+ messages in thread
From: uros at gcc dot gnu dot org @ 2010-04-19 12:37 UTC (permalink / raw)
To: gcc-bugs
------- Comment #5 from uros at gcc dot gnu dot org 2010-04-19 12:37 -------
Subject: Bug 43766
Author: uros
Date: Mon Apr 19 12:37:16 2010
New Revision: 158515
URL: http://gcc.gnu.org/viewcvs?root=gcc&view=rev&rev=158515
Log:
PR target/43766
* config/i386/i386.c (ix86_decompose_address): Handle ASHIFT addends.
testsuite/ChangeLog:
PR target/43766
* gcc.target/i386/pr43766.c: New test.
Added:
trunk/gcc/testsuite/gcc.target/i386/pr43766.c
Modified:
trunk/gcc/ChangeLog
trunk/gcc/config/i386/i386.c
trunk/gcc/testsuite/ChangeLog
--
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=43766
^ permalink raw reply [flat|nested] 7+ messages in thread
* [Bug target/43766] x86 prefetch doesn't use complex memory addressing
2010-04-16 7:52 [Bug target/43766] New: x86 prefetch doesn't use complex memory addressing astrange at ithinksw dot com
` (4 preceding siblings ...)
2010-04-19 12:37 ` uros at gcc dot gnu dot org
@ 2010-04-19 12:38 ` ubizjak at gmail dot com
5 siblings, 0 replies; 7+ messages in thread
From: ubizjak at gmail dot com @ 2010-04-19 12:38 UTC (permalink / raw)
To: gcc-bugs
------- Comment #6 from ubizjak at gmail dot com 2010-04-19 12:38 -------
Fixed.
--
ubizjak at gmail dot com changed:
What |Removed |Added
----------------------------------------------------------------------------
Status|ASSIGNED |RESOLVED
Resolution| |FIXED
Target Milestone|--- |4.6.0
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=43766
^ permalink raw reply [flat|nested] 7+ messages in thread
end of thread, other threads:[~2010-04-19 12:38 UTC | newest]
Thread overview: 7+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2010-04-16 7:52 [Bug target/43766] New: x86 prefetch doesn't use complex memory addressing astrange at ithinksw dot com
2010-04-16 11:37 ` [Bug target/43766] " ubizjak at gmail dot com
2010-04-16 18:07 ` ubizjak at gmail dot com
2010-04-16 21:19 ` astrange at ithinksw dot com
2010-04-17 15:11 ` ubizjak at gmail dot com
2010-04-19 12:37 ` uros at gcc dot gnu dot org
2010-04-19 12:38 ` ubizjak at gmail dot com
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for read-only IMAP folder(s) and NNTP newsgroup(s).